skipped bit levels - Page 6

Mark Rose · 2015-12-20, 21:38

Quote:

Originally Posted by alpertron

A small correction. At this rate, prime factors found by TJAOI will reach 2⁶⁰ at the end of December. Notice that on Fridays he submits prime factors of Mersenne number whose exponents are less than 1M found by ECM. These are obviously greater than that bound.

Good to know. I decided to do everything from bit 1 to what's in the history, just to be sure.

Mark Rose · 2015-12-20, 21:41

Quote:

Originally Posted by Madpoo

That's kind of cool... Kind of makes you think it could be worth some effort after all?

Well I haven't saved any DC work yet. I want to see how many more factors I find.

I'm doing everything where the history isn't recorded. There are a lot of exponents with no "no factors below 2^x" history lines.

flagrantflowers · 2015-12-22, 04:14

Quote:

Originally Posted by Mark Rose

Well I haven't saved any DC work yet. I want to see how many more factors I find.

I'm doing everything where the history isn't recorded. There are a lot of exponents with no "no factors below 2^x" history lines.

Please let us know if you find anything less than ~60 bits, would be really good to know if TJAOI missed anything.

Mark Rose · 2015-12-22, 06:13

Quote:

Originally Posted by flagrantflowers

Please let us know if you find anything less than ~60 bits, would be really good to know if TJAOI missed anything.

I plan to report all factors here.

Pretty much all the remaining work I have queued (35,000 exponents) has a history showing factoring up to 2^58 to 2^61 with a gap until 2^63. I should have that done by Monday.

Mark Rose · 2015-12-23, 16:20

An update for those curious:

I've checked all the skipped bit levels between 65M and 80M and found no factors.

Almost all the assignments above 80M lack a "no factor below 2^xx" history line. With these, I checked only the gaps between recorded history and found no factors.

I've also checked between everything between 35M and 51M with only that one factor found earlier.

52M to 54M had only two exponents; no factors.

I'm working on 51M to 52M and 54M to 65M right now.

chalsall · 2015-12-23, 16:46

Quote:

Originally Posted by Mark Rose

I'm working on 51M to 52M and 54M to 65M right now.

This is a valuable effort for those who wonder if factors have been missed; if people have been cheating.

I have done similar work in the past, and have never found a cheater.

"Many eyes make bugs shallow..."

Mark Rose · 2015-12-23, 17:30

Quote:

Originally Posted by chalsall

This is a valuable effort for those who wonder if factors have been missed; if people have been cheating.

I have done similar work in the past, and have never found a cheater.

"Many eyes make bugs shallow..."

I probably won't find any cheaters as I'm only looking where there is no history. If a cheater submits a result it will create an entry in the history.

When the work up to 65M is done, I may go look at the work below 35M. There's a few weeks of work there, too.

Madpoo · 2015-12-23, 18:18

Quote:

Originally Posted by Mark Rose

I probably won't find any cheaters as I'm only looking where there is no history. If a cheater submits a result it will create an entry in the history.

When the work up to 65M is done, I may go look at the work below 35M. There's a few weeks of work there, too.

FYI, Mark contacted me about taking his results from all of this and getting them added to the server so they show up in the history. I'm going to check with Mr. Prime95 and just make sure that's cool, but that way these gaps will be tidied up in the history of each exponent and we won't keep wondering about them in the future.

Prime95 · 2015-12-24, 03:45

I have no problem adding them to the history. I think Mark has proven that the old methods did properly track the bit levels even though we do not have a complete history.

Mark Rose · 2015-12-24, 04:36

Quote:

Originally Posted by Prime95

I have no problem adding them to the history. I think Mark has proven that the old methods did properly track the bit levels even though we do not have a complete history.

Seems like it! 25,000 exponent gaps tested so far and only one factor.

LaurV · 2015-12-24, 10:36

People are forgetting that, for a long time, before mfaktX reporting features were added to the server, we used to report only the factors, and not the "no factor" results that came out from the GPU pipes. Ask for example Mike (Xyzzy), he was one of the "biggest offenders" at that time, with a lot of GPU power, and only reporting factors, scrapping the "no factor" lines (and the credit, hehe, of course that was before he sold out his GPU farm). Therefore you should not waste your efforts to "fill in" the missing levels if a factor is already known, there are huge chances that the "skip" was done at that time, when we could not report mfaktc's "no factor" lines, but the work was done. Anyhow, if a factor is known, you don't help the project much by filling those gaps. This I was trying to say from the beginning.

No disrespect for your work intended, everyone does whatever he likes best to do, and double checking TF ranges (i.e. filling the gaps) may be useful for something in the future.

2015-12-24, 10:36	#66
LaurV Romulan Interpreter "name field" Jun 2011 Thailand 2×5,179 Posts	People are forgetting that, for a long time, before mfaktX reporting features were added to the server, we used to report only the factors, and not the "no factor" results that came out from the GPU pipes. Ask for example Mike (Xyzzy), he was one of the "biggest offenders" at that time, with a lot of GPU power, and only reporting factors, scrapping the "no factor" lines (and the credit, hehe, of course that was before he sold out his GPU farm). Therefore you should not waste your efforts to "fill in" the missing levels if a factor is already known, there are huge chances that the "skip" was done at that time, when we could not report mfaktc's "no factor" lines, but the work was done. Anyhow, if a factor is known, you don't help the project much by filling those gaps. This I was trying to say from the beginning. No disrespect for your work intended, everyone does whatever he likes best to do, and double checking TF ranges (i.e. filling the gaps) may be useful for something in the future. Last fiddled with by LaurV on 2015-12-24 at 10:40

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Normalising rent levels	Bundu	Math	4	2017-09-27 06:14
Racism or low light levels or...?	jasong	jasong	2	2016-09-25 05:07
Missing bit levels?	NBtarheel_33	Data	6	2016-05-31 15:27
Is the data missing or did we miss a couple TF bit levels	petrw1	PrimeNet	2	2015-05-07 05:09
Recommended TF bit levels for M(>10^8)	NBtarheel_33	Math	19	2008-11-03 17:19

2015-12-23, 16:20	#60
Mark Rose "/X\(‘-‘)/X\" Jan 2013 https://pedan.tech/ 6203₈ Posts	An update for those curious: I've checked all the skipped bit levels between 65M and 80M and found no factors. Almost all the assignments above 80M lack a "no factor below 2^xx" history line. With these, I checked only the gaps between recorded history and found no factors. I've also checked between everything between 35M and 51M with only that one factor found earlier. 52M to 54M had only two exponents; no factors. I'm working on 51M to 52M and 54M to 65M right now.

2015-12-24, 03:45	#64
Prime95 P90 years forever! Aug 2002 Yeehaw, FL 20246₈ Posts	I have no problem adding them to the history. I think Mark has proven that the old methods did properly track the bit levels even though we do not have a complete history.